| WwA: Worldwide Attitudes | [ WwA Home | Feedback | Search | International Survey Center Home] |
"The Effect of Parents’ Politics on their Children’s Education"
Jonathan Kelley
International Survey Center
(e-mail: jonathan.kelley@anu.edu.au)
(mail: ISP, RSSS, ANU/ Canberra ACT 0200/ Australia)
Revised version of my article in Worldwide Attitudes Vol. 1995-09-18:1-12
(c) Copyright 1997, Jonathan Kelley
Abstract
Political preferences are strongly linked to social class in many nations, especially Anglo-Celtic ones. Whatever their own class position, supporters of conservative parties are inclined to sympathize with established social institutions like schools and employers and inclined to stress individual’s responsibility and control over their own fate; supporters of left-wing parties often take the opposite views. I suggest that these views have implications for children’s education: that supporters of conservative parties are -- regardless of their own class position -- more likely to encourage their children to stay on in school, to work hard in school, and eventually to obtain high status occupations. Conversely, supporters of left-wing parties will be less appreciative of education; less inclined to encourage individual effort; and more satisfied with working-class occupations for their children.
To investigate this possibility empirically, I analyze data from a pooled sample of six surveys conducted between 1984 and 1995 by the International Social Science Survey / Australia with 13,642 respondents. The analysis controls for all the ‘usual suspects’ -- parents’ education, father’s occupation, number of books in the home, number of siblings, gender, urbanization, historical period, and family structure.
Taking all this into account, conservative parents get their children about half a year more education than Labor parents get their children (t=9.6, p < .0001). To put this in perspective, it is approximately equivalent to the educational advantage of being born into a family with two university educated parents, rather than two parents who only completed secondary school. The advantage comes about mainly because conservative families are more successful in getting their children to complete secondary school: for example, if a child from a family where both parents are Labor supporters has a 40% chance of finishing secondary school, a child from a conservative family with exactly the same education, occupation, and other characteristics would have a 52% chance of finishing.
A similar analysis of the 1979 British Election Survey (N=1,855) leads to very similar results: Other things being equal, Conservative families in Britain get their children 0.7 years more education than Labour families get for their children.
Political preferences are strongly linked to social class in many nations, especially Anglo-Celtic ones, although more so in the past than in the present (Lipset 1983; Kelley and Evans 1995: Kelley, McAllister and Mughan, 1985; Nieuwbeerta, 1995). Whatever their own objective circumstances, supporters of conservative parties are inclined to sympathise with higher classes rather than lower classes; inclined to favour management rather than workers; inclined to sympathise with established social institutions like big business, churches, and schools; and inclined to stress individual’s responsibility and control over their own fate. Conversely supporters of left-wing parties are inclined to view the lower classes favourably; to favour workers rather than management; to distrust established social institutions; and to attribute success or failure to society rather than to individual effort and ability. Such views have many and diverse implications for how people perceive the world around them and for politics.
I suggest that these views also have implications for the family, particularly for the children’s education. Because they identify with the upper classes, that supporters of conservative parties are -- regardless of their own class position -- more likely to want their children to get into upper class occupations. Because they trust schools, conservatives are more likely to encourage their children to stay on in school beyond the legal minimum, and to work hard in school. Because conservatives see individuals as responsible for their own fate, they are more likely to think their children can and should succeed by dint of effort and persistence in school, and to encourage them to do so. Conversely, supporters of left-wing parties will be more satisfied with working-class occupaitons for their children; less appreciative of education and schools; and less inclined to encourage individual effort and diligence.
To investigate this possibility empirically, I analyze data from a pooled sample of six surveys conducted between 1984 and 1995 by the International Social Science Survey / Australia with 13,642 respondents (Evans and Kelley, 1995). I also replicate the analysis in Britain using the 1979 British Election Survey.
Children whose parents both support the conservative Liberal-National coalition get, on average, 11.5 years of education (table 1). Fully 47% finish secondary school (year 12) and 18% finish a university degree. In sharp contrast, children from families that support the Labor Party get a year less education, 10.4 years. Only 28% finish secondary school and just 9% get a university degree. Thus parents’ politics seems to affect their children’s education.
However, this takes no account of other differences between conservative parents and parents who support the Labor Party. Conservative parents tend themselves to come from the higher classes with higher status jobs and all the attendant advantages. They are also somewhat more scholastically inclined, with more education, more books in the home, and sometimes a history of academically rigorous private schooling themselves. All these provide educational advantages to their children, quite apart from politics. In contrast, Labor families are less favourably situated to help their children educationally, quite apart from politics. To see whether politics itself matters, we must take these other differences into account.
Many aspects of class and family background shape a child’s education. In this Australia is very similar to other nations, with the ‘usual suspects’ (table 2) influencing education in the usual way (table 3; see Teachman 1987 for similar results in the United States):
- Having parents who both finished a 3-year university course (the usual length in Australia) instead of parents who only finished secondary school adds an extra 0.6 year of schooling, other things being equal.
- Having parents who went to private school -- although rare -- also helps, adding another 0.6 years of education. Note that this is an effect of private schooling per se, and not because these parents are well educated themselves, or hold elite jobs (the analysis adjusts for those factors).
- Having parents who went to Catholic schools also helps, by about the same amount.
- Having parents with scholarly inclinations, as reflected in a large number of books about the house, helps a lot. In this Australia is like many other nations (De Graaf 1986; DiMaggio 1982).
Taking everything else into account, conservative parents get their children about half a year more education than Labor parents get their children. This difference is statistically highly significant (t=9.6, p < .0001). To put it in perspective, it is approximately:
This conclusion is not an artifact of the particular model estimates. Alternative models lead to similar conclusions (for example, table 4, line 2).
The advantage comes about mainly because conservative families are more successful in getting their children to complete secondary school. The advantage is about 12% (table 4, line 4). For example, if a child from a family where both parents are Labor supporters has a 40% chance of finishing secondary school, a child from a conservative family with exactly the same education, occupation, and other characteristics would have a 52% chance of finishing.
Children from conservative families are about 5% more likely to finish university than are children from Labor families, other things being equal. However, this is, only because they are more likely to have finished secondary school and so become eligible for university. A secondary school graduate from a conservative family has no better chance of going on to finish university than an otherwise similar secondary school graduate from a Labor family (table 4, line 4).
Australian politics was more strongly linked to class in the past than it was in the 1980s and subsequently (Bean and Kelley 1988; Nieuwbeerta 1995: 49-52, 70-74, 111-119). Presumably because of that, parents politics also had more of an educational impact in the past than they do now (table 4, line 3):
At that rate, class differences would entirely disappear by the end of the century.
These results are not unique to Australia. A similar analysis of Britain leads to very similar results (table 5). Other things being equal, Conservative families in Britain get their children 0.7 years more education than Labour families get for their children.
At least in Australia and Britain -- two of the more class-conscious Anglo-American nations -- it seems reasonably clear that politically conservative parents get their children a little further on in school than Labor parents do. This is net of parents objective class position. One danger (suggested by a perceptive reviewer) is that parents’ subjective class is not controlled: it is possible that subjective class shapes both parents’ politics and their children’s educational careers. Moreover even if the effect of parents’ political views is real, the mechanisms through which it influences education remain highly speculative. I argued that it has to do with conservative parents attaching greater value to education and the success it brings in the class system; trusting schools and other institutions more than Labor parents do; and seeing success as more surely attainable through individual effort rather than Labor parents do. But this argument needs testing: the intervening mechanisms should be measured and the argument demonstrated with evidence rather than merely asserted with rhetoric. Unfortunately, we have no empirical measures of these things for parents. Moreover asking respondents to report (say) the subjective class of their parents might not provide very reliable data anyway (parents’ education, occupation and party are known to be reliably reported; parents’ income is poorly reported; and reports of parents’ attitudes and values are generally quite unreliable). So the theoretical argument remains untested in this analysis.
However we do have measures for subjective class and other relevant values for respondents themselves. So we could turn the analysis about: instead of asking how respondents’ education is affected by their family background (the conventional approach, which I follow in this paper), we could take respondents as the ‘parents’ and ask how their (well measured) values and objective class position affect their children’s education. This requires data on children’s education, which is available in some of our surveys. In a future paper, I hope to report the results of such an analysis.
Australia
The International Social Science Survey / Australia is a nation-wide survey conducted by researchers at the Australian National University and the University of Melbourne. Begun in 1984, it is Australia's leading academic survey and is devoted entirely to academic research in the social sciences, is non-profit, and is not connected with any business or political party. The survey's core sponsor is the Research School of Social Sciences at ANU. The ISSS is based on large, representative national samples of all states and territories, drawn from the electoral roll. The detailed and comprehensive survey takes about two hours to complete. It is conducted by mail. The first mailing includes a cover letter from the Australian National University and a postage-paid reply envelope, followed by a further letter about two weeks later. Anyone who did not respond within a month or so is then pursued by up to three more mailings over a six month period. Comparison with the census shows the samples collected in this way to be representative of the Australian population in age, sex, education, occupation, and other characteristics. Dr. Jonathan Kelley is Director and principal investigator of the ISSS; Dr. Clive Bean (Associate Director), Dr. M.D.R. Evans and Dr. Krzysztof Zagorski are co- principal investigators.
This research reports on data from a pooled sample of six surveys from 1984 to 1995. There are 14,034 respondents of whom 13,642 provided information on education.
Britain
The British data are from the well-known 1979 British Election Survey. There are 1,893 cases of whom 1,855 provided information on education..
Australian Bureau of Statistics. 1983. "Census ‘81: Educational Qualifications" Order no. 2149.0. Canberra: Australian Government Publishing Service.
Bean, Clive and Jonathan Kelley. 1988. "Partisan Stability and Short-term Change in the 1987 Federal Election: Evidence from the NSSS Panel Survey". Politics: The Journal of the Australian Political Studies Association. 23: 80-94.
De Graaf, P. M. 1986 "The impact of financial and cultural resources on educational attainment in the Netherlands" Sociology of Education 59 (4): 237-246.
DiMaggio, P. 1982 "Cultural capital and school success" American Sociological Review 47 (2): 189- 201.
Evans, M.D.R. and Jonathan Kelley. 1995. Education in Australia: International Social Science Survey, Pooled File, 1985-1995. Codebook and Machine Readable Data File. Canberra: International Social Science Survey, Institute of Advanced Studies, Australian National University.
Evans, M.D.R., Jonathan Kelley, M. Borgers, J. Dronkers, and L. Rollenberg. 1995. "Parental Divorce and Children's Education: Australian Evidence" Worldwide Attitudes 19950717: 1-8.
Kelley, Jonathan and M.D.R Evans. 1995 "Class and class conflict in six Western nations" American Sociological Review 60 (April): 157-178.
Kelley, Jonathan, Ian McAllister and Anthony Mughan. 1985. "The Decline of Class Revisited: Class and Party in England, 1964-1979" American Political Science Review. 79: 719-737
Lipset, S. M. 1983. Political Man: The Social Bases of Politics. (expanded and updated edition) London: Heinemann.
Nieuwbeerta, Paul. 1995. The Democratic Class Struggle in Twenty Countries, 1945/1990. Amsterdam: Thesis Publishers.
Teachman, Jay D. "Family Background, Educational Resources, and Educational Attainment" American Sociological Review 52: 548-557.
Table 1. Description: Parents’ politics and their child’s education.
Australia, 1984-1995.
------------------------------------------------------------------
Years of Finish Finish
education secondary university
(mean) (percent) (percent) Cases
------------------------------------------------------------------
Parents both support
conservative parties 11.5 47% 18% (4151)
Parents views
mixed, or unknown 10.8 39% 13% (4758)
Parents both Labor
Party supporters 10.4 28% 9% (4733)
-----------------------------------------
Total 10.9 38% 13% (13,642)
------------------------------------------------------------------
Note: The conservative parties are the Liberal and the National
(formerly Country) Parties, long-time coalition partners.
The (very few) supporters of centrist parties are included
in the mixed category.
Table 2. Variable means and standard deviations.
--------------------------------------------------------------------------
Mean Std Dev Description
--------------------------------------------------------------------------
Dependent variables:
Education 10.89 2.86 Years of education (based on a detailed
questions on (i) years of primary &
secondary schooling and (ii) highest degree,
diploma or certificate (ABS 1983)scored
according to usual length of time required
for completion)
Year12 .38 .48 Completed year 12 (=1), or not (=0)
University .13 .34 Completed university (=1), or not (=0)
Parents party .48 .41 Parents party. Conservative parties are
scored 1, centrist parties and parents whose
politics are unknown to their child scored
0.5, and Labor supporters scored 0. If
only one parent’s politics is known, that
is used; if both are known, they are
averaged.
Parents education:
Years 8.80 2.65 Parents education (mean)
Private schl .07 .21 Parents attended private, 2ndry school (mean)
Catholic schl .13 .29 Parents attended Catholic 2ndry school (mean)
# of books 138.46 244.62 # books in parents home when R age 14
Fathers occupation:
Status 38.68 24.96 Fathers occupational status (0 to 100)
Not know .09 .28 Fathers occupation not known
Controls:
# of siblings 3.09 2.06 Number of siblings
Year born 44.57 16.31 Year born (yrs since 1900)
Male .50 .50 Male (=1), Female (=0)
PntDvrcd .06 .24 Parent divorced (when R age 14)
PntDead .08 .27 One parent dead (when R age 14)
Mother worked .26 .37 Mother worked for pay when R was age 0-14
Quadratics:
PntEducSq 8.47 14.66 Quadratic: Pnt Ed - 10, squared
FaStatSq 624.55 913.40 Quadratic: FaStat3 - 40, squared
BooksSq 61315.83 243629.08 Quadratic: # of books - 100, squared
BornSq 295.44 417.95 YearBorn - 50, squared
Interaction:
PPtyXAge 21.36 21.01 Interaction: parents party x Year born
--------------------------------------------------------------------------
Table 3: Influences on educational attainment: Australia, 1984-1995.
Ordinary least squares regression predicting years of education.
N = 13,642.
----------------------------------------------------------------
Variable B SE B Beta T Sig T
----------------------------------------------------------------
Parents' party
(conservative=1) .54 .06 .08 9.59 .00
Parents' education:
Years .19 .01 .18 19.32 .00
Private school .58 .11 .04 5.32 .00
Catholic school .45 .08 .04 5.70 .00
# books at home .001 .00 .09 11.17 .00
Father's occupation:
Status .02 .00 .15 16.69 .00
Occ not known -.32 .08 -.03 -3.87 .00
Controls:
YearBorn .05 .00 .29 34.92 .00
# of Sibs -.17 .01 -.12 -15.56 .00
Male .34 .04 .06 7.79 .00
Pnts divorced -.58 .09 -.05 -6.13 .00
Pnt dead -.20 .09 -.02 -2.39 .02
Mother worked .05 .06 .01 .79 .43
Constant 6.20 .11 57.48 .00
Adjusted R Square .31
Standard Error 2.37
----------------------------------------------------------------
Table 4. Alternative models: Effects of parents' political party
preference on respondent's education. Australia, 1985-1995 [a]
----------------------------------------------------------------------
Slope Standard Signi-
(metric) error ficance
----------------------------------------------------------------------
Predicting years of education
-----------------------------
1. Basic OLS model (Table 3) .54 .06 p < .0001
2. Model 1 + quadratics for parents'
education, # books, father's status,
and respondent's year of birth .48 .06 p < .0001
3. Changes over time: Predicted slope
from model 2 + interaction with cohort [b]
(a) In secondary school around 1925 .84 --
(b) In secondary school around 1955 .53 --
(c) In secondary school around 1985 .22 --
Predicting educational transitions [c]
--------------------------------------
4. Logistic regression, variables from
models 1 & 2 predicting probability of:
(a) Completing secondary school .12 .01 p < .01
(b) Completing university, if
completed secondary school .04 .02 not sig.
----------------------------------------------------------------------
Notes:
[a] N = 13,642 cases for the dependent variable in OLS, treating
missing data by the pairwise present method. Logistic regression uses
only cases with complete information on all variables, giving
N = 11,785 for the analysis of secondary completion and N = 3,628
for university completion.
[b] The interaction coefficient for (parents party x year of birth)
is -.010 with standard error of .003 (p<.0001). We assume respondents
are in secondary school around age 15.
[c] Slopes, evaluated at the mean of the dependent variable.
Table 5. : Influences on educational attainment: Britain, 1979.
Ordinary least squares regression predicting years of education.
N = 1,855.
----------------------------------------------------------------
Variable B SE B Beta T Sig T
----------------------------------------------------------------
Parents' party
(conservative=1) .74 .11 .14 6.66 .00
Father's occupation:
Status .03 .00 .32 13.69 .00
Owner -.05 .14 -.01 -.33 .75
Farmer .45 .16 .07 2.91 .00
Controls:
Age -.04 .00 -.38 -18.89 .00
Male .14 .08 .03 1.75 .08
Non-white .80 .25 .06 3.25 .00
Catholic .15 .13 .02 1.15 .25
Lives in south .18 .08 .04 2.19 .03
Constant 11.17 .15 75.26 .00
Adjusted R Square .30
Standard Error 1.70
----------------------------------------------------------------
[a] Parents’ education was not asked. It is only modestly
correlated with parents’ party (Kelley et al, 1985), so its
omission will probably give only a slight upward bias to the
parents’ party coefficient.
Visitors since 14 September 1998: